Vehicle Detection in Aerial Surveillance Using Dynamic Bayesian Networks

Vehicle Detection in Aerial Surveillance Using

Dynamic Bayesian NetworksHsu-Yung Cheng, Member, IEEE, Chih-Chia

Weng, and Yi-Ying ChenIEEE TRANSACTIONS ON IMAGE

PROCESSING, VOL. 21, NO. 4, APRIL 2012

Goal

Introduction

• These technologies have a variety of applications, such as military,police, and traffic management.

• Aerial surveillance is more suitable for monitoring fast-moving targets and covers a much larger spatial area.

Introduction

• Cheng and Butler [8] performed color segmentation via mean-shift algorithm and motion analysis via change detection.

• In [11], the authors proposed a moving-vehicle detection method based on cascade classifiers.

• Choi and Yang [12] proposed a vehicle detection algorithm using the symmetric property of car shapes.

Introduction

Background Color Removal

• quantize the color histogram bins as 16*16*16.

• Colors corresponding to the first eight highest bins are regarded as background colors and removed from the scene.

Feature Extraction:Local Feature Analysis

Feature Extraction:Local Feature Analysis

Feature Extraction:Local Feature Analysis

• After evaluation, is known.

• Use the gradient magnitude G(x,y) of each pixel of moment-preserving.

• Tmax =T ,Tmin=0.1*(Gmax-Gmin) for Canny edge detector.

• Harris detector is for the corners.

Feature Extraction:Color Transform and Color Classification

• In [16],they proposed a color domain (u,v) instead of (R,G,B) to separate vehicle and non-vehicle pixels clearily.

• Use n*m as a block to train SVM model to classify color.

Feature Extraction:Color Transform and Color Classification

Feature Extraction

• We extract five types of features, S,C,E,A and Z for the pixel.

• A=L/W• Z=blue counts at left

Dynamic Bayesian Network

• Use some videos to train the probabilities with people marked ground truth.

• Vt indicates if a pixel belongs to a vehicle.

• P(Vt|St) is defined as the probability that a pixel belongs to a vehicle pixel at time slice given observation St at time Instance t.

Experimental results

Experimental results

Experimental results

Experimental results

Experimental results

Experimental results

Experimental results